https://github.com/tom-doerr/intrinsic-image-popularity
https://github.com/tom-doerr/intrinsic-image-popularity
Last synced: 7 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/tom-doerr/intrinsic-image-popularity
- Owner: tom-doerr
- Created: 2021-01-03T17:04:13.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2022-06-08T02:09:22.000Z (about 4 years ago)
- Last Synced: 2025-02-14T09:48:11.908Z (over 1 year ago)
- Language: Shell
- Size: 204 MB
- Stars: 1
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Intrinsic image popularity Assessment
This is a PyTorch implementation of the paper [Intrinsic Image Popularity Assessment](https://arxiv.org/abs/1907.01985).
This work can quantitatively predict how much traction an image will get on Instagram.
It can help users to find the image that would be the most popular to the public.
Run ```python test.py --image_path ``` to evaluate the intrinsic image popularity of your photos on Instagram.
This is the [Online Demo](https://iipa.ngrok2.xiaomiqiu.cn) (unstable).
### Dataset
We provide the dataset of popularity-discriminable image pairs by the form of "shortcode". You can download the images with the URL ```"https://www.instagram.com/p//media/?size=l".```
*Note, some URLs may be invalid now.*
### Citation
```
@inproceedings{ding2019intrinsic,
title={Intrinsic Image Popularity Assessment},
author={Ding, Keyan and Ma, Kede and Wang, Shiqi},
booktitle={ACM International Conference on Multimedia},
pages={1979--1987},
year={2019},
publisher={ACM}
}
```